pull down to refresh

Why would it solve the IBD problem? AFAICT, IBD bottlenecks are primarily hardware related.

In my limited understanding, Core's utxoset design results in a lot of linear processes, whereas the libbitcoin model (on which Rearden based this I think) is much more easily parallelized.

I think when downloading blocks, you can check all sorts of things before you check the truly linear thing (do these inputs exist and have they been spent before). Core doesn't necessarily do all the optimization in can here because it is based on the design of Satoshi's original client.

reply

Yeah, that seems plausible to me. I had major IBD slowness a few years ago and wrote about it on SN. I think I isolated it down to either a slow disk speed (I had a SSD but connected via a USB3 cable), or slow CPU (I was running it on a Rpi4). Possibly an issue with RAM speeds. But I never really isolated the problem. I'm pretty sure it wasn't data bandwidth or amount of RAM though.

reply

I think USB speed was the culprit. After starting an IBD on pi4, I took out SSD and plugged it into my gaming PC booted to Debian from a thumb drive. That IBD finished in a few hours if I remember correctly.

reply